The aim of the current study was to evaluate a multiplex PCR (mPCR) detection test combined with the evaluation of a previously described isolation method. Minced beef, raw-milk cheese and sprouted seed samples were inoculated with low amounts (7-58 cfu 25 g(-1)) of non-stressed, cold-stressed or freeze-stressed clinical STEC strains, including serogroups O26, O103, O111, O145, sorbitol fermenting (SF) O157 and non-sorbitol fermenting (NSF) O157. The inoculated pathogen was detected using a 24 h-enrichment followed by an mPCR protocol, and in parallel isolated using an enrichment step of 6 and 24 h, followed by selective plating of the enriched broth and selective plating of the immunomagnetic separation (IMS) product. Recovery results were evaluated and compared. Successful mPCR detection and isolation was obtained for non-stressed and cold-stressed STEC cells in minced beef and raw-milk cheese samples, except for serogroups O111 and SF O157. For freeze-stressed cells and sprouted seed samples, false negatives were often found. Isolation was better after 24 h-enrichment compared to 6 h-enrichment. IMS improved in some cases the isolation of non-stressed and cold-stressed cells belonging to serogroups O111 and O157 from minced beef and raw-milk cheese and freeze-stressed cells of all tested serogroups from minced beef.